Good manufacturing practices production of natural killer cells for immunotherapy: a six-year single-institution experience.

BACKGROUND: Natural killer (NK) cells, a subset of lymphocytes and part of the innate immune system, play a crucial role in defense against cancer and viral infection. Herein is a report on the experience of clinical-scale, good manufacturing practices (GMPs) production of NK cells to treat advanced cancer. STUDY DESIGN AND METHODS: Two types of NK cell enrichments were performed on nonmobilized peripheral blood mononuclear cell apheresis collections with a cell selection system (CliniMACS, Miltenyi): CD3 cell depletion to enrich for NK cells and CD3 cell depletion followed by CD56 cell selection to obtain a more pure NK cell product. After overnight incubation with interleukin-2 (IL-2), cells were washed, resuspended in 5 percent human serum albumin, and then released for infusion. RESULTS: A total of 70 NK cell therapy products have been manufactured for patient infusion since 2000. For the CD3 cell-depleted NK cell products, the mean purity, recovery, and viability were 38, 79, and 86 percent, respectively. For the CD3 cell-depleted/CD56 cell-enriched NK cell products, the mean purity, recovery, and viability were 90, 19, and 85 percent, respectively. Gram stain, sterility, and endotoxin testing were all within acceptable limits for established lot release. Compared to the resting processed cells, IL-2 activation significantly increased the function of cells in cytotoxicity assays. CONCLUSION: Clinical-scale production of NK cells is efficient and can be performed under GMPs. The purified NK cell product results in high NK cell purity with minimal contamination by T cells, monocytes, and B cells, but it requires more time for processing and results in a lower NK cell recovery when compared to NK cell enrichment with CD3 cell depletion alone. Additional laboratory studies and results from clinical trials will identify the best source and type of NK cell product.

Recapitulation of oral mucosal tissues in long-term organotypic culture.

To test the influence of fibroblasts on epithelial morphology and expression of keratinocyte proteins and barrier lipids, we bioengineered homotypic and heterotypic oral mucosae and skin using cultured adult human cells. Fibroblasts were allowed to modify collagen type I gels for 2 weeks before keratinocytes were added. The organotypic cultures were then grown at the air-liquid interface for 4 weeks. In homotypic combinations, epithelial morphology and protein expression closely mimicked those in vivo. In heterotypic combinations, the morphology resembled that in vivo and keratinocytes expressed their typical markers, except when skin keratinocytes were recombined with alveolar fibroblasts; they expressed K19, K4, and K13, which is similar to oral mucosal epithelia rather than to the epidermis. Morphologically, the stratum corneum layers were typical for the epithelial tissues. Grafting the bioengineered cultures to the backs of Nude mice did not change the results, suggesting that our findings are not merely a culture phenomenon. Lipid profiles of the homotypic combinations mimicked the profiles found in the normal epithelial tissues, except that the engineered alveolar epithelium expressed more ceramide 2 than that in vivo. In the heterotypic combinations, keratinocytes appeared to control the lipid profile, except in the combination of skin keratinocytes with alveolar fibroblasts, wherein the ceramide profile appeared to be partly that of alveolar epithelium and partly that of epidermis. These results suggest that cultured adult fibroblasts and keratinocytes are sufficient to recapitulate graftable oral tissues, and, except for alveolar fibroblasts, the type of fibroblast had little influence on keratinocyte differentiation.